TheDataGirl

Linear Regression (also known as linear association) is often used in data science and is known as a basic method of predictive analysis as it can easily determine the relationship between two or more independent variables. This method is ideal when the two variables’ relationship is linear, that is, that the relationship between them can be depicted as a straight line. There are two types of linear regression, simple and multiple. Simple Linear Regression is when the relationship is mapped between two variables; the outcome variable and the predictor variable. Multiple Linear Regression is when the relationship is mapped between more than two variables. One variable is the outcome variable and the other variables are the predictor variables.

Linear regression can be defined by the equation,

y=β_1+β_2 X+ ∈

y is the outcome variable also known as a response variable
β_1 is the intercept
β_2 is the slope
β_1 and β_2 are known as regression coefficients
X is the predictor variable. This is also known as the explanatory variable
∈ is the error
(Prabhakaran)

Advantages of Linear Regression:

• It is very simple and easy to use
  Its simplicity makes it an obvious choice for many data scientists and has been reported by KD Nuggets as the popular choice in 2016 and 2017. (Montero Porras, 2018)
• Allows for a margin of error
• Very fast

Disadvantages of Linear Regression:

• Does not work well with non-linear relationships
• Results can be greatly affected by outliers
• Data must be independent
  (Flom, 2018)

References
Flom, P. (2018, March 13). The Disadvantages of Linear Regression. Retrieved from Sciencing: https://sciencing.com/disadvantages-linear-regression-8562780.html
Montero Porras, E. (2018, July 18). Linear Regression in R. Retrieved from Data Camp: https://www.datacamp.com/community/tutorials/linear-regression-R
Prabhakaran, S. (n.d.). Linear Regression. Retrieved from R-statistics: http://r-statistics.co/Linear-Regression.html